One-Shot Fault Diagnosis of Wind Turbines Based on Meta-Analogical Momentum Contrast Learning

The rapid development of artificial intelligence offers more opportunities for intelligent mechanical diagnosis. Fault diagnosis wind turbines is beneficial to improve the reliability turbines. Due various reasons, such as difficulty in obtaining fault data, random changes operating conditions, or compound faults, many deep learning algorithms show poor performance. When samples are small, ordinary will fall into overfitting. Few-shot can effectively solve problem overfitting caused by fewer samples. A novel method based on meta-analogical momentum contrast (MA-MOCO) proposed this paper very few turbine failures, especially one-shot. By improving (MOCO) and using training idea meta-learning, one-shot drivetrain analyzed. model shows a higher accuracy than other common models (e.g., model-agnostic meta-learning Siamese net) learning. feature embedding visualized t-distributed stochastic neighbor (t-SNE) order test effectiveness model.